generated: '2026-10-03'
method: generated
source: three first-party SKILL.md files served at https://humanmirror.fr/skills/<name>/SKILL.md (added 2026-10-03), plus openapi/ (operationIds verified against the specs; Nexus/Outcome specs carry none, so those steps cite METHOD
  /path), mcp/humanmirror-fr-nexus-tools.json, and the provider's own skill.md which was searched and saved verbatim.
skills:
- file: humanmirror-fr-genesis-skill.md
  name: humanmirror-genesis
  method: searched
  source: https://humanmirror.fr/skill.md
  description: 'Provider-published Agent Skill: machine-native deterministic context cleanup with a free Genesis
    bootstrap (20 calls / 48h) and x402 paid continuation, plus the Machine Port, Pulse, Gateway, Next-Gen preview,
    Escrow v1, Shared Context, Perception signals and Intent Market entry points.'
  operations: null
  note: Saved verbatim; it is prose + endpoints rather than operationIds, and it contains the provider's zero-width
    provenance watermark.
- file: humanmirror-fr-secure-external-context.md
  name: humanmirror-secure-external-context
  method: generated
  description: Quarantine prompt injection, scan secrets, preflight the tool call and bind an execution receipt
    over x402.
  operations: 5
- file: humanmirror-fr-verified-outcome.md
  name: humanmirror-verified-outcome
  method: generated
  description: Trial key, free quote, success-only run and free re-verification through Nexus/Outcome (REST and
    MCP).
  operations: 6
- file: humanmirror-fr-purchase-firewall.md
  name: humanmirror-purchase-firewall
  method: searched
  first_party: true
  source: https://humanmirror.fr/skills/purchase-firewall/SKILL.md
  original: skills/_original/purchase-firewall-SKILL.md
  description: 'Provider-published Agent Skill: bind an autonomous purchase or payment intent before execution (POST /api/value/bind/),
    let the native payment rail execute, then reconcile the observed result (POST /api/value/close/) into MATCH / DRIFT evidence.'
  operations: null
  note: Verbatim copy in skills/_original/; the skills/ copy changes only the frontmatter name to the Agent Skills slug form (provider name kept as x-provider-name) and adds api/operations/source. The provider announced it at github.com/VibeMirror-coder/humanmirror/skills (404 on 2026-10-03, repo
    not public); the same file is served at the source URL. /api/value/bind/ and /api/value/close/ answer live (400 on an empty body) but
    are not described in any published OpenAPI, so no operationIds are listed.
- file: humanmirror-fr-safe-preflight.md
  name: humanmirror-safe-preflight
  method: searched
  first_party: true
  source: https://humanmirror.fr/skills/safe-preflight/SKILL.md
  original: skills/_original/safe-preflight-SKILL.md
  description: 'Provider-published Agent Skill: inspect a proposed agent command or tool action before execution and return an
    ALLOW / REVIEW / BLOCK-style decision via the x402-paid safe-preflight endpoint (0.010 USDC on Base).'
  operation_count: 1
  operations:
  - humanmirrorX402SafePreflight
  note: Verbatim copy in skills/_original/ (frontmatter name slugged in the skills/ copy). Endpoint answered 402 with a PAYMENT-REQUIRED header on 2026-10-03.
- file: humanmirror-fr-execution-receipt.md
  name: humanmirror-execution-receipt
  method: searched
  first_party: true
  source: https://humanmirror.fr/skills/execution-receipt/SKILL.md
  original: skills/_original/execution-receipt-SKILL.md
  description: 'Provider-published Agent Skill: submit expected and observed action context after a consequential action and store
    the returned deterministic execution receipt (x402-paid, 0.010 USDC on Base).'
  operation_count: 1
  operations:
  - humanmirrorX402ExecutionReceipt
  note: Verbatim copy in skills/_original/ (frontmatter name slugged in the skills/ copy). Endpoint answered 402 with a PAYMENT-REQUIRED header on 2026-10-03.
exclusions:
- surface: Physical Oracle worker registration / payout
  reason: A2H field-work flow involving human workers' payout addresses; not packaged as a skill pending review
    of the personal-data implications.